https://securingdev.github.io/posts/ai-security-reasoning-and-bias/